Journal of Experimental Marine Biology and Ecology 462 (2015) 62–73
Contents lists available at ScienceDirect
Journal of Experimental Marine Biology and Ecology
j ourna l homepage: www.e lsev ie r .com/ locate / jembe
Interactions between the invasive European green crab, Carcinus maenas(L.), and juveniles of the soft-shell clam, Mya arenaria L., in easternMaine, USA
Eric Bryan P. Tan, Brian F. Beal ⁎Division of Environmental and Biological Sciences, University of Maine at Machias, 116 O'Brien Ave., Machias, ME 04654, USA
Invasive species pose a threat to biodiversity in numerous marine ecosystems, and may have severe economiceffects on commercially important species. The European green crab, Carcinus maenas, is one of the most com-mon invaders of marine ecosystems globally. Since its invasion into eastern Maine, USA, during the early1950's, populations of the soft-shell clam, Mya arenaria, have declined greatly. This has triggered the establish-ment of shellfish hatcheries and the development of aquaculture techniques to enhance the wild fishery. Thisstudy investigated interactions between C.maenas and cultured juveniles ofM. arenaria both in the field and lab-oratory. In the field (Holmes Bay, Cutler, Maine), clam (initial mean shell length [SL] ±95% CI: 15.8 ± 0.5 mm;n=30) survival was: 1) 7× higher in predator deterrent treatments compared to open controls; 2) not improvedby using rigid vs. flexible netting; and, 3) not improved by raising and supporting deterrent netting 5 cm abovethe sediment surface. Wild clam recruitment was 4x greater in protected vs. open experimental units. In labora-tory trials using similar sized juvenile clams, green crabs consumed clams protected by predator deterrentnetting, and in one case did so without leaving visible signs of chipping, crushing, or disarticulating the valves.Physical evidence, other than crushing, may be used to differentiate between clam death due to predation vs.suffocation, disease, or other sources of mortality.
Predation has played a key role in limiting population growth ofmarine organisms in the evolutionary past (Sallan et al., 2011; Stanley,2008), and continues to drive important direct and indirect effectson populations in modern marine ecosystems (Babcock et al., 1999;Freestone et al., 2011; Nakaoka, 2000). Studies examining in situ inter-actions between decapod predators and their infaunal bivalve prey(Irlandi, 1997; Kuhlmann and Hines, 2005; Seitz et al., 2001; Whitlow,2010; Wong et al., 2010) have shown the importance of interactive fac-tors such as prey andpredator size, prey and predator behavior, bottom/habitat type, crab molt frequency, and large-scale environmentalperturbations such as hypoxia (Seitz et al., 2003) and other stressors(Smee et al., 2010) to help explain bivalve mortality patterns.
Soft-shell clams,Mya arenaria L., are ubiquitous residents of the soft-bottom intertidal and shallow subtidal zone in the northeast (Beukema,1976) and northwest Atlantic (Conde et al., 2010; Hunt, 2004), wherethey are also an important commercial species (Beal, 2002; Dow,1977). Mya feeds by filtering phytoplankton from the water column
through its long, fused siphons, and burrows deeply with age to avoidpredation (Zaklan and Ydenberg, 1997). Behavioral responses such asincreasing burial depth (Flynn and Smee, 2010;Whitlow, 2010) and re-duced growth (Beal et al., 2001) occur in the presence of predators. Ju-veniles (b15 mm shell length, SL) of M. arenaria live at or near thesediment-water interface (LeBlanc and Miron, 2006), and during thisearly part of its life history crustaceans (Bowen and Hunt, 2009; HuntandMullineaux, 2002; Taylor and Eggleston, 2000) and other predatorssuch as fish (Kelso, 1979; Steimle et al., 2000) may nip the siphons, orremove individuals completely from the sediments to consume them(Blundon and Kennedy, 1982; Smith et al., 1999). As Mya increases insize, it becomes prey to infaunal predators such as naticid gastropods(Edwards and Hubner, 1977), nemertean worms (Bourque et al.,2001), and other species that are adept at removing it from sedimentsand consuming it at the surface such as large decapod crustaceans(Floyd and Williams, 2004; Ropes, 1968; Seitz et al., 2001; Smith andHines, 1991). Mya may reach a size-refuge (Commito, 1982) or spatialrefuge (Skilleter, 1994) from some predators, although these may notbe generalizable across the clam's geographic distribution (Beal,2006a). In Maine, USA, and other New England states, cyclical declinesin soft-shell clam populations since the mid-1900's have been largelyassociated with the dynamics of the invasive green crab, Carcinusmaenas (Cohen et al., 1995; Whitlow, 2010).
63E.B.P. Tan, B.F. Beal / Journal of Experimental Marine Biology and Ecology 462 (2015) 62–73
Short-term field experiments in easternMaine over the past two de-cadeswith juveniles ofM. arenaria have used flexible netting as ameansto deter predators to enhance bivalve survival (Beal, 2006a,b; Beal andKraus, 2002; Beal et al., 2001). Most trials have shown the importanceof predation in regulating clam populations, with the use of exclusionnetting explaining between 5 and 45% of the variation in clam survival;however, relatively high percentages of missing and/or dead clamswith crushed valves also have been associated with predator exclusiontreatments in these experiments (e.g., up to 34.6% see Beal, 2006b).The cause of missing and crushed clams from experimental unitscompletely protected with netting, and with no predator (live ordead) present (or the presence of molts or molted appendages) in ex-perimental units at the end of the field trial, has not been establishedunambiguously. Chipping of the ventral margin of dead clams ofM. arenaria and other bivalves, or nearly complete crushed valves(articulated or disarticulated) has been used as the primary physicalevidence related to decapod predation (Beal, 2006a,b; Boulding andLabarbera, 1986; Peterson, 1982) while empty valves that show nosigns of chipping or crushing have been attributed either to naturaldeath, disease such as neoplasia (Weinberg et al., 1997), death due tosuffocation, or death from benthic, nemertean worms that leave nomark after consuming their prey (Bourque et al., 2002).
Here, alternative types of predator netting are tested in the field andlaboratory on survival of juveniles of Mya arenaria to help explain thefate ofmissing and crushed individuals from previous exclusion studies,and examine interactions and effects of the presence of green crabs,Carcinus maenas, on clam survival through a series of laboratory exper-iments. Green crabs were used because they have been observed withincreasing frequency in easternMaine over the past decade: 1) in inter-tidal soft-bottoms (Beal, pers., obs.); and, 2) in experimental units filledwith ambient sediments designed to exclude this and other predators(Beal, pers. obs.). In addition, C.maenas have becomemore common re-cently elsewhere in Maine within intertidal soft bottoms (ME DMR,2013; Whitlow and Grabowski, 2012). The working hypothesis is thatin previousfield trials (described above), successful attacks by C.maenason completely protected clams occurredwhen crabs rested on top of theflexible netting, weighing it down, and then used the tips of their chelaethrough the apertures (4.2 mm or 6.4 mm) to excavate bivalves fromthe sediment by grasping their siphons. Clams were then consumedeither through the netting, or on the outside of the netting when crabsforced small pieces of shell and tissue through the netting apertures.Further, an additional hypothesis is that missing clams would havebeen crushed to an extent that any remaining umbos and shellfragments in experimental units at the end of the field trial wouldhave been smaller than the aperture of the sieves (1–2 mm) used towash the sediments from each unit (see Beal et al., 2001 for a descrip-tion of these methods). If the prediction is correct, then treatmentsusing rigid netting or flexible netting that is supported and raisedsufficiently above the sediments of experimental units containingjuveniles of soft-shell clams, or units partially filled with sediment andcovered with netting, should not contain crushed or missing clamsafter sufficient time has elapsed for predators to access clams.
Table 1Description of each of the five levels of the exclusion factor from Field Experiment I (15 July to 2was crossed with both types of netting (Flexible or Rigid; 6.4 mm aperture). (n=5; N=50).
Exclusion Description Abbrev
Predator Exclusion with Netting Unsupported PENUPredator Exclusion with Netting Supported 5 cm Off Bottom With Metal Rods PENSPredator Exclusion with Netting Unsupported & Metal Rods Present as a Control PENROpen Enclosure with Peripheral Mesh Collar to Deter Clam Migration OEMCOpen Enclosure with Supported Roof Control OERC
2. Methods
2.1. Field Experiment I. (15 July–29 October 2011; Duck Brook Flat, Cutler,Maine)
This experiment was conducted in unvegetated sediments near thelower mid-intertidal at Duck Brook Flat (DBF) in Holmes Bay, Cutler,Maine, USA (44° 41′ 39.04″ N, 67° 18′ 17.82″ W), from 15 July to 29October 2011 (106 days; see Beal and Kraus, 2002 for a description ofthis site). Experimental units (plastic horticultural pots, 0.018 m2 andca. 15 cm deep), similar to those used by Beal et al. (2001), were usedto test the interactive effects of netting type (a=2) and exclusion treat-ment (b=5; Table 1; Fig. 1) on soft-shell clam survival and growth.Each type of netting was fully crossed with each exclusion treatmentyielding ten treatments. Predator exclusion netting was either rigid(extruded) or flexible with a 6.4 mm aperture (Industrial Netting; Min-neapolis, MN; http://www.industrialnetting.com/). The five exclusiontreatments were: 1) fully protected with netting that was supportedca. 5 cm from the sediment surface using four short pieces of metalwire rods fabricated from coat hangers (2.5 mm diameter; PredatorExclusion Netting Supported – PENS); 2) fully protected with nettingthat was unsupported (no metal rods; Predator Exclusion NettingUnsupported – PENU); 3) fully protected with unsupported netting,but with metal rods present to control for potential artifacts caused bythe presence of the rods (Predator Exlusion Netting unsupported, butwith Rods present – PENR); 4) open enclosure with a 1-cm rim of net-ting around the outside periphery to deter clammigration andwith fourshortmetal rods (Open Enclosurewith peripheralMesh Collar –OEMC);and, 5) open enclosure without a peripheral collar, but with a piece ofnetting to serve as a roof control that allowed predators access to theexperimental units, but with a circular piece of netting (ca. 15 cmdiameter) supported by wire rods ca. 5 cm from the sediment surface(Open Enclosure Roof Control – OERC).
Metal rods used in eight of the ten treatments had hooked ends(Fig. 1), and were affixed to the units by heating up the straight end,piercing it horizontally through the unit about 1 cm below the rim,and then bending it in such a way to secure it to the pot. Four rodswere evenly spaced along the circumference of the rim of each unit,and were measured beforehand to ensure a fixed distance betweenthe rim of the pot and the top of the hooked end. Units fully protectedwith pieces (50 cm × 50 cm) of unsupported, flexible netting (Fig. 1-G)were similar to those used in past experiments at this site (Beal andKraus, 2002). That is, the netting completely covered the top of theunit and was raised a small distance above the surface of the sediments(by pulling the netting upward after installing the unit in the sediments)to prevent the netting from interfering with clam feeding. It is possible,however, that the slackness in the netting associatedwith this treatmentis enough to allow predators such as crabs to rest on the top of the net-ting, weighing it down sufficiently so that it might be possible for themto access clams through the apertures. To examine this hypothesismore closely, a second, fully protected treatment was used except theflexible netting was raised approximately 5 cm above the sediment and
9October 2011) at Duck Brook Flat in Holmes Bay, Cutler, Maine, USA. Each exclusion type
Features
iation Netting Present Predator Deterrent Metal Rods Netting Supportedby Metal Rods
Fig. 1. Photos taken in the field showing each of the 10 treatments used in Field Experiment I at Duck Brook Flat (DBF) in Holmes Bay, Cutler, Maine, USA. The top row (A-E) shows eachtreatment with rigid netting, and the second row (F–J) shows each treatment with flexible netting (experimental units=15 cm diameter; aperture size=6.4 mm). (See Table 1 for a de-scription of abbreviations used for each exclusion treatment.) Open enclosures (A & F: OEMC) had a rim, or collar, of netting around the periphery to deter clammigration, plus fourmetalwires affixed to the inside periphery at regular intervals. Full netting treatments (B & G: PENU), typical of normal exclusion treatments used in other field studies (e.g., Beal and Kraus,2002); full netting treatments (C & H: PENS) with the top of the net supported ca. 5 cm from the sediment surface with metal wires to prevent netting from collapsing under the weightof crustacean predators; full netting treatments (D & I: PENR), similar to B & G, but that control for the presence of metal wires; and roof controls (E & J: OERC) that control for the possibleeffects of the netting roof (shading, reduced flow, increased sedimentation, etc.) while allowing predators access to clams in the units.
64 E.B.P. Tan, B.F. Beal / Journal of Experimental Marine Biology and Ecology 462 (2015) 62–73
supported by securing it to four metal rods using cable ties (Fig. 1-H).Pieces of both flexible and extruded netting (26 cm × 2 cm) used forthe fully open treatments were wrapped around the outer periphery ofexperimental units, and were held in place using heavy-duty rubberbands (Fig. 1-A & F). Pieces of extruded netting used for the three fullyprotected treatments (Fig. 1-B–D) were cut, shaped, and held togetherwith nylon cable ties so that they would fit tightly and over each experi-mental unit once it was placed in the sediments. The circular pieces ofextruded and flexible netting (Fig. 1-E & J) were affixed to the fourmetal rods in each unit using nylon cable ties.
On 15 July 2011, a 5 × 10matrix was established, and each of the tentreatments assigned randomly to positions within each row (completelyrandomized block design; CRBD). Adjacent experimental units in bothrows and columnswere placed approximately 1m apart to ensure inde-pendence of treatments. Experimental units were filled with ambient,muddy sediments, and then pushed into the sediments so that a smalllip (ca. 0.15 cm) extended above the sediment surface. Next, six juvenilesof Mya arenaria (mean shell length ±95% CI=15.8 ± 0.5 mm; n=30),produced (cultured) in 2010 at the Downeast Institute for Applied Ma-rine Research & Education (DEI; Beals, ME) and overwintered accordingto Beal et al. (1995), were pushed gently into the top cm of sediment ineach unit. This resulted in an initial stocking density of 330 individualsm−2. Prior to the establishment of the experimental matrix, 10 benthiccores were taken from the experimental area using a coring device(area=0.0182 m2 to a depth of 15 cm) to establish ambient densitiesof wild clams and green crabs. Samples were transported to theUniversity ofMaine atMachias (UMM)where eachwaswashed througha 2 mm sieve.
All experimental units were retrieved after 106 days on 29 October2011, and taken to UMM where the contents of each was washedthrough a 2 mm sieve. All live, dead undamaged (DU), and deadchipped/crushed (DC) hatchery-reared individuals of Mya arenaria,Carcinus maenas, and juvenile wild recruits of Mya arenaria (SL≤15 mm) were collected and enumerated. Any crushed valve withoutan umbo was disregarded and not collected. Missing clams also werenoted. A growth rate was estimated for each live experimental clambymeasuring its initial SL using a distinct disturbance line incorporatedin both valves at the time when cultured clams are transferred to sedi-ments (see Beal et al., 1999). Both initial SL and final SL were measuredto the nearest 0.01 mm using digital calipers. A mean relative growthindex ([Final SL – Initial SL]/Initial SL)was obtained for each experimen-tal unit containing at least one live clam.
2.2. Field Experiment II. (3 July–28 October 2011; Duck Brook Flat,Cutler, Maine)
To enhance information about the ability of predators to prey onjuveniles of soft-shell clams protected by flexible netting, a second fieldexperiment was initiated on 3 July 2011 at DBF in unvegetated soft sedi-ments near themid intertidal (within ca. 50mof Experiment I). The effectof two interactive factors on clam survival were tested: intraspecific clamdensity (a=5; 3, 6, 12, 24, or 48 clams per unit representing densities of165, 330, 660, 1320, and 2640 individuals m−2, respectively), and preda-tor exclusion (b=2; flexible, unsupported exclusion netting similar toPENU [see above], but with 4.2 mm aperture vs. open enclosures similarto OEMC [see above], but without metal rods and with the peripherysurrounded by a 1 cm piece of flexible netting [4.2 mm aperture]to deter clam migration). Experimental units (as described above)were filled with ambient sediments and arranged in a 5 × 10 matrixwith 1-m spacing between rows and columns. Replicates of the ten treat-ments were randomly assigned a position within the matrix. Culturedjuveniles of M. arenaria (mean SL ±95% CI=13.0 ± 0.7 mm; n=40),with the same origin and history as those used in Experiment I, wereadded to experimental units as described above. All units were collectedfrom DBF after 117 days on 28 October 2011, and processed at UMM asdescribed above.
2.3. Laboratory Experiment I (21 August to 11 September 2011; DEI,Beals, ME)
A laboratory experiment was conducted at the Marine EducationCenter at DEI on Great Wass Island, Beals, Maine, USA (44° 28 ′ 50.64″N, 67° 35′ 54.97″ W) from 21 August to 11 September 2011 (22 days)to examine closely the behavior of green crabs preying on soft-shellclam juveniles held in treatments similar to those used in Field Experi-ment I (seawater temperature ranged from 13–15 °C). Six clams, pro-duced at DEI in 2011 (mean shell length ±95% CI: 17.5 ± 0.7 mm,n=49), were added to plastic horticultural pots (as described above)filled with a commercial sand (Sakrete® Natural Play Sand, Charlotte,North Carolina). Three pots were added to each of three 75-l aquaria(clear glass sides) and then sediment was added around each pot suchthat it completely encased each pot to its rim. One pot in each aquariumwas covered with a piece of rigid or flexible netting (similar to PENU –
Table 1) while the other pot was covered with a tight-fitting piece offlexible netting that would not bend or depress under the weight of a
65E.B.P. Tan, B.F. Beal / Journal of Experimental Marine Biology and Ecology 462 (2015) 62–73
crab. Unfiltered, ambient sea water flowed into each aquarium at alltimes during the 22-day trial period (13–15 °C). One male green crab(carapace width, CW=60.2 mm, 64.9 mm, and 61.9 mm; each had re-cently completed ecdysis, and each had a hard carapace) was added toeach aquarium at the beginning of the trial. Crabs were collected fromthe rocky intertidal adjacent to DEI the day before the trial. During thefirst eight days, crabs would sometimes climb out of the aquaria over-night. At that time, a wooden board was used to cover the top of eachaquarium to ensure that crabs would remain confined to the aquarium.
Daily observations of each aquarium were made to estimate thenumber of live clams (by counting the number of visible siphons atthe sediment surface) in each treatment, number of dead or crushedclams at the sediment surface, the position of the crab within the tank,and the percent of time each crab spent burrowed. At the end of the ex-periment, the contents of each pot waswashed through a 1.8mm sieve,and the number of live, dead undamaged (DU), and dead chipped/crushed (DC) clams recorded.
2.4. Laboratory Experiment II (3–9 January 2012; UMM, Machias, ME)
A second laboratory experiment was conducted in the aquacultureresearch laboratory at UMM from 3 to 9 January 2012. This experimentwas designed to determine whether the physical presence of greencrabs or chemical cues from this species would alter feeding activity ofcultured juvenile soft-shell clams. To understand the fate of microalgaewithout any grazers and in the absence of light, prior to the beginning ofthe actual experiment (27 December 2011), four experimental units(diameter=17 cm, depth=15 cm) were filled with one liter of sand(as described above), a total of 1.5 l of seawater, and approximately50,000 cells mL−1 of a microalgal monoculture (Chaetoceros gracilis[CCMP 1316]). Seawater used in the trial was collected at DEI, andfiltered through a 1 μmfilter bag. Each experimental unit received an in-dependent supply of air. The first estimate of algal cell numbers in eachexperimental unit was conducted the next day, and again after a periodof 6 days, on 2 January 2012, when the final count was recorded. All cellcounts were completed using a haemocytometer. Some water hadevaporated from the four units during the period of time; therefore,the volume of water left in each unit wasmeasured and counts adjustedto obtain a density per 1.5 L. Sediment in each unit was checked foranoxia. None had an odor, and the sediment in each had not turnedblack indicating that the level of aeration was sufficient for a 6-dayperiod in the experimental units.
Room temperature during the pre-trial period and throughout thesecond laboratory experiment was maintained between 15–17 °C, asthis was approximately the temperature range in Laboratory Experi-ment I. During the pre-trial period and throughout Laboratory Experi-ment II, the room was kept in complete darkness, and lights wereturned on only when data was being recorded. The experimentwas ini-tiated on 3 January 2012 and ran for 6 days. Sixteen experimental units(as in the pilot study, and filled to 2 L with filtered seawater – 34 ppt)were arranged into four blocks, and each of four treatments wasassigned randomly to a position (2× 2matrix) within each block. Treat-ments were: 1) Control (C) –microalgae only (5 × 104 cells mL−1; thistreatment was similar to the pilot experiment); 2) Microalgae pluscrushed green crab extract (CC) and six hatchery-reared juvenile soft-shell clams (mean SL ±95% CI=19.3 ± 0.6 mm, n=51; produced atDEI during 2011); 3) Microalgae plus one live male crab (LC) and sixhatchery-reared juvenile soft-shell clams; and, 4) Clam control (CL) –
microalgae plus six hatchery-reared clams only. Crabs were obtainedon 3 January 2012 from the rocky intertidal area adjacent to DEI, andeach experimental unit was aerated independently. Because it was dif-ficult to capture crabs at this time of year, crabs of both sexes were in-cluded in the trial; however, only males (CW range=26.5–37.9 mm)were used in the LC treatment. Two of the four green crabs used in theCC treatment were females (CW=27.7 mm and 38.6 mm). Clamswere pushed gently into the sediment as in previous experiments. For
experimental units associated with the CC treatment, each crab wascrushed in a glass dish and ca. 10 ml of orange extract poured into aunit ca. 30 minutes after clams were established in the sediments. Asmall amount of seawater in each of the sixteen units was replaced onfive of the seven days during the trial by mixing filtered seawater withaged freshwater kept in a HDPE carboy. To determine effects of treat-ment on the fate of algal cells, counts were recorded on Days 3 and 6using a haemocytometer, and each measurement was an average ofthree counts from each treatment. At the end of the experiment, clamswere collected by sieving the sediments from each experimental unitusing a 1 mm sieve. Number of live, DU, and DC clams from each unitwas recorded.
2.5. Laboratory Experiment III (27 June–17 July 2013; DEI, Beals, ME)
To further explore the ability of green crabs to consume juveniles ofMya arenaria protected with flexible netting, a third laboratory experi-ment was conducted. Plastic horticultural plant pots (as describedabove) were added in groups of three to each of five 75-l aquariums(T=18–20o C). Sand (as described above)was added to each potwithineach aquarium to create three treatments as follows: pot #1 – to therim; pot #2 – within 1 cm of the rim; and pot #3 – within 8 cm of therim. Next, sand was added to each aquarium so that the three experi-mental units were buried (on the outside of each unit) to their rim.Six clams (mean SL=14.8 ± 0.4 mm, n=30) were placed on the sedi-ment surface of each unit and then a piece of flexible netting (6.4 mmaperture) with slack, similar to that used in Field Experiment I(i.e., PENU) and Laboratory Experiment I, was secured to each potusing a rubber band. One male individual of Carcinus maenas(CW=69.9–71.7 mm) was added to each aquarium after clams hadcompletely burrowed. Crabs were collected from the rocky shoreadjacent to DEI, and were allowed 24 hr to acclimate to experimentaltemperatures. Seawater (32 ppt) in each aquarium was changed twiceduring the experimental interval, and clams were fed daily 100 ml ofcultured T. Isochrysis galbana. The prediction from field trials was thatif green crabs were able to consume soft-shell clams through themesh netting, that clams would be consumed only in units filled tothe rim with sand. At the end of the experiment, number of live anddead (DU, DC) clams in each experimental unit was noted.
2.6. Data analysis
For Field Experiment I, the relative efficiency of the CRBD(Underwood, 1997) against a completely randomized design (CRD)was lower for each dependent variable. Hence, blocking was ignoredand the data analyzed as a CRD using analysis of variance (ANOVA),and the following linear model:
Yijk ¼ μ þ Ai þ Bj þ ABij þ ek ijð Þ
Where:
Yijk thedependent variable (meanpercent survival,mean relativegrowth,mean number of recruits, andmean number of greencrabs);
Ai Netting type (j=1–2, fixed);Bj Enclosure type (i=1–5, fixed); and,ek Experimental error (k=1–5).
Relative growth (RG= [Final SL− Initial SL]/Initial SL) was used toassess growth rather than final SL or absolute growth because of thesignificant linear relationship between initial and final SL (r2=0.128,P b0.0001, n=141), which was similar to previous studies (Beal,2006b). An RG=1 indicates a doubling of shell length. An arcsine trans-formation was performed on mean percent survival for the analysis ofvariance (ANOVA) to meet assumptions of variance homogeneity and
66 E.B.P. Tan, B.F. Beal / Journal of Experimental Marine Biology and Ecology 462 (2015) 62–73
normality. No transformations were necessary for other dependentvariables.
A series of mutually exclusive, single degree-of-freedom orthogonalcontrasts was performed to more closely examine the nature of thepredator exclusion treatments:
1) Predator Exclusion Netting vs. Controls (PENU, PENR, PENS vs.OEMC, OERC). This tests the overall importance of predation.
2) Predator Exclusion Netting (Unsupported and unsupported withmetal rods present) vs. Predator Exclusion with supported Netting(PENU, PENR vs. PENS). This tests the importance of keeping nettingabove the sediment surface to ensure crabs and other predatorscannot gain access to the clams through the netting.
3) Predator Exclusion Netting (Unsupported) vs. Predator ExclusionNetting (Unsupported, but with metal rods present (PENU vs.PENR). This tests whether the presence of the metal rods affectsany of the dependent variables.
4) Migration control vs. Roof controls (OEMC vs. OERC). For the openenclosures, this tests whether the presence of a roof (but without aperipheral collar to deter migration) affects clam survival, growth,recruitment, or the presence of green crabs differently than inunits without a roof but with a peripheral fence (collar).
Because one treatment associated with the relative growth data(roof control with flexible netting) had no survivors among the fivereplicates, and some experimental units from the remaining nine treat-ments contained no live clams, the planned contrasts were not orthog-onal to each other. To control for the lack of independence among theplanned contrasts, an adjusted type I error rate was used according toWiner et al. (1991):α′=1− (1-α)1/n (where n=number of contrasts;α′=0.0127).
The linear model used for the analysis of survival data from FieldExperiment II, which were arcsine-transformed to meet assumptionsof ANOVA, was identical to that used for Field Experiment I.
For Laboratory Experiment I, the frequency of the crab at one of fourpositions within each aquarium (1: beside treatment 1 (tight-fitting,flexible netting); 2: beside treatment 2 (slack-fitting, flexible netting);3: beside treatment 3 (rigid netting); and, 4: between pots)was record-ed, and a Fisher's Exact Test performed to determine if crab behavior dif-fered between aquaria. To determine if siphonal activity was related tocrab position within the aquaria, the number of visible siphon pairs inthe pot next to the crab was recorded each day for each aquarium.This value was subtracted from the mean number of siphon pairsin the other two pots in the aquarium on a particular day. Themean difference should be zero under the null hypothesis that proxim-ity of a crab to a particular pot does not affect siphon activity. A one-sample t-test on the mean difference was performed to test this nullhypothesis.
Data from Laboratory Experiment II was analyzed as a CRD usingANOVA because of the lack of a significant block effect (P N0.75). Thelinear model was the same as that used for Field Experiment I, where:
Yijk is the dependent variable (untransformed mean number ofcells mL−1);
Ai Date (i=1–2, fixed);Bj Treatment (j=1–4, fixed); and,ek Experimental error (k=1–4).
A series of single degree-of-freedom orthogonal contrasts was usedto examine more closely the treatment effect:
1) Crab vs. No Crab (LC & CC vs. C & CL);2) Live Crab vs. Crushed Crab (LC vs. CC); and,3) Control (without clams) vs. Control (with live clams) (C vs. CL).
Data from Laboratory Experiment III was analyzed as a randomizedcomplete block design, with one replicate of each treatment per block
(=aquarium). The dependent variable was number of dead, crushed(DC) clams per unit.
All analyses were conducted using SAS 9.2. Sample means are pre-sented with their corresponding 95% confidence intervals.
3. Results
3.1. Field Experiment I. (15 July–29 October 2011; Duck Brook flat, Cutler,Maine)
Mean number of wild clams from the benthic cores taken on 15 Julyfrom the experimental area prior to establishing the trial was 10.9 ±16.6 individuals m−2 (n=10). Two clams (13.1 mm and 19.4 mm) oc-curred in two of the ten samples. None of the core samples contained agreen crab or other large, potential crab predators (e.g., Cancer irroratus,B. Beal, pers. obs.).
3.1.1. SurvivalFew missing and dead clams with crushed/chipped valves were
observed from fully protected units (PENU, PENS, PENR) regardless ofnetting type (Table 2). Unlike previous field trials, combined lossesdue to missing and dead crushed/chipped valves from fully protectedunits with unsupported flexible netting (PENU) was only 3.3 ± 9.3%(a single clam in one of the five experimental units was found deadwith distinct chipping along its ventral margin; Table 2). This was theonly fully protected treatment where any dead clams with crushed orchipped valves was found; however, a single clam was missing fromone of the five units covered with unsupported flexible netting andwith the metal rods extending up through the netting (PENR) and inone unit with rigid netting supported 5 cm off the sediment surface bymetal rods (PENS; Table 2). Mean percent of missing and crushed/chipped (DC) clams from the open enclosures (OEMC, OERC) rangedfrom 56.7% to 86.7%; concomitantly, the presence of netting explainednearly 71% of the total variation in mean survival (Table 3). No signifi-cant difference in mean survival was observed between treatmentswhere rigid (48.0 ± 14.9%, n=25) vs. flexible netting (46.0 ± 14.6%,n=25) was used (P=0.4118), and the pattern of survival across exclu-sion treatments was similar between netting types (P=0.9283; Fig. 2).Among exclusion treatments, the orthogonal contrast that examinedthe importance of predation (fully protected units [PENU, PENS, PENR;70.6 ± 8.3%, n=30] vs. open controls [OEMC, OERC; 11.7 ± 8.0%,n=20]) explained nearly 92% of the variation (P b0.0001). The onlyother significant contrast was among the controls (P=0.0089), wheremean survival among units with roof controls (pooling netting type)was 1.7± 3.8% (n=10) vs. 21.7 ± 13.8% (n=10) for units with collars.None of the clams in the roof controls with flexible netting (OERC;Fig. 1–J) survived (Table 2), suggesting that the peripheral collarsprevented clams from leaving the pots, collars reduced predator effec-tiveness, or predator effectiveness was enhanced somehow due to thepresence of a roof.
3.1.2. GrowthRelative growth did not vary significantly either with exclusion type
or netting (Table 4; Fig. 3). Mean final SL pooled across all treatmentswas 22.4 ± 0.7 mm (n=38; Table 5), an absolute increase of 5.9 ±0.7 mm.
3.1.3. Wild Mya recruits and Carcinus maenasWild recruits of Mya occurred in 36 of 50 (72%) experimental units
(range=1–17 individuals unit−1, or 55–935 ind. m−2). Mean numberof recruits was ca. four times higher in units fully protectedwith netting(4.6 ± 1.4 ind., n=30) compared to controls (1.1 ± 0.5 ind., n=20;Fig. 4), and this difference was statistically significant (P=0.0003;Table 6). In addition, significantly more wild recruits occurred in unitswith netting supported by metal rods (PENS: 6.2 ± 3.4 ind., n=10)vs. those with unsupported netting (PENU, PENR: 3.8 ± 1.4, n=20;
Table 2Mean (±95% CI) percent alive (%A), dead with undamaged valves (%DU), dead with crushed or chipped valves (%DC), and missing (%M) of hatchery-reared juveniles of Mya arenaria,mean number of wild recruits ofMya (SL ≤15 mm), and mean number of green crabs, Carcinus maenas from experimental units (Area=0.0182 m2) associated with Field ExperimentI at Duck Brook Flat, in Holmes Bay, Cutler, Maine (15 July to 29 October 2011). See Table 1 for a description of each level of the Exclusion factor, and Fig. 1 for a photo of each. (n =5).
Netting Exclusion % A % DU % DC % M Recruits Crabs
67E.B.P. Tan, B.F. Beal / Journal of Experimental Marine Biology and Ecology 462 (2015) 62–73
Table 6). Individuals of C.maenas (CW range=6.7 mm to 13.7 mm) oc-curred in six units (10%), all of which were protected with netting(P=0.0433; Table 6). Although five of the six units were protectedwith rigid vs. flexible netting, this result was not statistically significant(P=0.0754; Table 6). A single crab was found in five units, and twocrabs were found in one unit. None of the units with green crabscontained DC clams.
3.2. Field Experiment II (3 July–28 October 2011; Duck Brook Flat,Cutler, Maine)
The effect of stocking density on juvenile soft-shell clam survivalwas not statistically significant (P=0.2913; Table 7); however, thenetting treatment explained nearly 53% of the total variation inclam survival (P b0.0001; mean clam survival in fully protectedunits=66.5±8.8% vs. 14.8± 10.1% in open units, n=25), and this pat-tern was similar across all densities (P=0.3628, Table 7). Clams werefound dead with chipped/crushed valves or were completely missingfrom each of the five density treatments with predator netting(Table 8). A combined mean of 20% of individuals were missing orcrushed from the highest density treatment. A total of 16 live greencrabs was found in 10 of the 50 (20%) experimental units (units had 1,2, or 5 crabs; 3 units were open enclosures, 7 units had protectivenetting). No significant effect due to the netting or density treatmenton mean green crab number was observed (P N 0.09). Fifteen crabshad CW less than 9 mm (CW range=3.8 mm to 8.7 mm; meanCW=6.1 ± 1.0 mm), and none of the units with these crabs containedcrushed clams. One male green crab (CW=30.0 mm) was found in anopen unit (initial density=24 clams/unit). No live clamswere observedin that unit and only three crushed valves were recovered, the remainingclams were missing. No significant effects due to density or nettingtreatment were observed on mean number of wild recruits ofM. arenaria (P N 0.30; x=1.4 ± 0.45 ind. unit−1, or 74.7 ± 24.6 ind.m−2; n=50).
Table 3Analysis of variance on the arcsine-transformed mean percent survival of hatchery-reared juvMaine (15 July to 29October 2011). Two factors (Plastic netting [a=2;Rigid or Flexible; aperturfactor]) were orthogonal to each other. The 4 single degree-of-freedom orthogonal contrasts fohypothesis tests (α=0.05). (n=5).
Source
ExclusionPredator Exclusion (PENU, PENS, PENR) vs. Open Controls (OEMC, OERC)Predator Exclusion Netting Unsupported (PENU, PENR) vs. Supported (PENS)Predator Exclusion Netting Unsupported (with vs. without metal rods: PENU vs. PENR)Open Enclosures (with collars vs. roof controls: OEMC vs. OERC)
Netting (Rigid vs. Flexible)Exclusion × NettingErrorTotal
3.3. Laboratory Experiment I (21 August to 11 September 2011; DEI,Beals, ME)
Each crab settled close to one of the pots more than 50% of the time(i.e., N11 daily observations; Table 9); however, the pattern of crab be-havior was not consistent between aquaria (P=0.0006). Two crabs (inaquaria “A” and “B”) behaved similarly, spending the majority of time(ca. 55%) adjacent to units protected with rigid netting, while thethird crab spent the majority of time burrowed (ca. 77%) and adjacentto the pot with tight-fitting, flexible netting. On day eight (30 August),a dead, articulated shell with no tissue was observed in aquarium “C”(treatment=slack, flexible netting) on top of the sediment and under-neath the netting. The clam had been chipped on its posterior ventralmargin, in the area where siphons are located. In addition, the flexiblenetting was closer to the sediment than it had been on previousdays, suggesting the crab had climbed on top of the netting, forcing itdown toward the sediment surface during the time it preyed on theclam. On one occasion, a crab was found resting on top of rigid netting(3 September 2011; aquarium “A”). This pot had two visible pairs ofsiphons while the other two pots in the same tank had 4 visible pairof siphons. In general, however, number of visible pairs of siphons wasnot related to the position of the crab within the aquarium (T=0.36,P=0.717, n=63). On 11 September, a dead clam with undamagedvalves, and some tissue remaining, was found in a pot in tank A withrigid netting. All other clams were found alive.
3.4. Laboratory Experiment II (3–9 January 2012; UMM, Machias, ME)
Mean algal cell densities differed significantly among treatments(P b 0.0001; Table 10). Orthogonal contrasts demonstrated that treatmentswith either a crushed crab (CC) or a live crab (LC) had cell counts fourtimes higher than those without crabs (17.63 ± 4.42 × 104 cells mL−1 vs.4.81 ± 1.52 × 104 cells mL−1, n=4) suggesting an effect on clam feed-ing. This result was consistent between sampling dates (P=0.3294;Table 10; Fig. 5). Mean algal cell densities from both crab treatments
eniles of Mya arenaria from Field Experiment I at Duck Brook Flat, in Holmes Bay, Cutler,e=6.4 mm] andExclusion [b=5; see Table 1 for a description of each level of the Exclusionr the Exclusion factor are presented. P-values in boldface represent statistically significant
Fig. 2. Untransformed mean percent survival (+95% CI) from Field Experiment Iconducted at DBF (15 July to 29 October 2011). (See Table 1 for a description of eachlevel of the Exclusion treatment, and Fig. 1 for a photo of each treatment). (n=5).
Exclusion Treatment
Mea
n R
elat
ive
Gro
wth
+ 9
5% C
I
0.0
0.2
0.4
0.6
0.8
1.0
FlexibleRigid
PENU PENR PENS OERC OEMC
Fig. 3. Untransformed mean relative growth (+95% CI) from the field experimentconducted at DBF (15 July to 29 October 2011). (See Table 1 for a description of eachlevel of the Exclusion treatment, and Table 5 for the number of replicates associatedwith each bar.).
68 E.B.P. Tan, B.F. Beal / Journal of Experimental Marine Biology and Ecology 462 (2015) 62–73
(CC vs. LC, P=0.3573) and both controls (C vs. CL, P=0.2607)were notsignificantly different (Table 10).
At the end of the experiment, clam survival in controls (CC& CL)wasapproximately 94% (3 of 48 animals were found dead, with undamagedvalves; Table 11). As with Laboratory Experiment I, the three DU clamsfrom the two treatments without crabs were found below the sedimentsurface, had blackened shells with a distinct odor typical of H2S, andmost of the decomposed tissue remained within the valves. Of thefour units containing clams and live crabs (LC), all clams were foundalive in one unit (that crab had burrowed into the sediments and wasmissing a chela), and all clamswere dead in the other three units; how-ever, clam fate varied greatly between units. Animals in two of the LCunits were all found deadwithin 24 hours after the experimentwas ini-tiated with chipped or crushed valves whereas those in the other unithad no discernible valve damage and were scattered on the surface ofthe sand within 24 hours. None of the valves (crushed or undamaged)in these three LC units had any attached flesh (i.e., shell and shell frag-ments were clean) indicating that crabs had consumed the soft tissueof each clam.
3.5. Laboratory Experiment III (27 June–17 July 2013; DEI, Beals, ME)
Crabs preyed on clams only in experimental units that werecompletely filled with sediments (P=0.0067; Table 12). All dead
Table 4Analysis of variance on the untransformed mean relative growth of live hatchery-rearedjuveniles of Mya arenaria from Field Experiment I at Duck Brook Flat, in Holmes Bay,Cutler, Maine (15 July to 29 October 2011). (See Table 1 for levels of these factors, andTable 3 for a description of each single degree-of-freedoma priori contrast associatedwiththe Exclusion factor.) Because no survivors occurred in one of the treatments (OERC-Flexible; See Table 2), a priori contrasts were not orthogonal; therefore, an adjusted typeI error rate was used (α′=0.0127). (n ranged from 1 to 5; therefore, Type III sums ofsquares were used in the analysis – see Shaw and Mitchell-Olds, 1993).
Source df SS MS F Pr N F
Exclusion 4 0.0674 0.0169 1.05 0.3985Fully Netted vs. Controls 1 0.0476 0.0476 2.97 0.0957Unsupported vs. Supported Netting 1 0.0025 0.0025 0.15 0.6986Unsupported (with vs. without metal rods) 1 0.0058 0.0058 0.36 0.5540Collar vs. Roof Control 1 0.0617 0.0617 3.85 0.0594
clams (n=5) were recovered as crushed valves with only the umbosrecognizable, and remaining relatively intact.
4. Discussion
Field and laboratory studies presented here indicate how vulnerablejuveniles of the soft-shell clam are to crustacean attack, especially greencrabs, Carcinus maenas, supporting results from previous investigationsfrom the northeast U.S. and Atlantic Canada (Beal et al., 2001; Elner,1981; Floyd and Williams, 2004; Hunt, 2004; Pickering and Quijón,2011; Whitlow, 2010). Field studies examining the role of predationon the fate of juveniles (b15mmSL) of the soft-shell clam,Mya arenaria(Beal, 2006a,b; Beal et al., 2001) have noted dead chipped and crushedclams as well as completely missing individuals in treatments fullyprotected by netting. Similar observations have beenmade for juvenilesof other infaunal bivalves such as Mercenaria mercenaria and/or Chionecancellata (Beal, 1983; Nakaoka, 2000; Peterson, 1982), Katelysiascalarina and K. rhytiphora (Peterson and Black, 1993), and when inves-tigators used netting in mariculture research programs (Beal and Kraus,2002; Cigarría and Fernández, 2000; Serdar et al., 2007; Spencer et al.,1992). Several hypotheses may explain the apparent mystery ofcrushed and missing clams from fully protected experimental units inthe field. 1) Chipped or crushed clams could result from unintentional,improper handling of small bivalves by investigators when initiatingthefield studies, or at the endof thefield trial during sample processing;2) Small predators could have been added accidentally to experimental
Table 5Mean relative growth index and final SL of hatchery-rearedMya arenaria (±95% CI) fromexperimental units associated with Field Experiment I at Duck Brook Flat, in Holmes Bay,Cutler, Maine (15 July to 29 October 2011). n=number of experimental units containinglive clams. (Mean initial SL ±95% CI=15.8 ± 0.5 mm; N=30.) See Table 1 for a descrip-tion of each level of the Exclusion factor, and Fig. 1 for a photo of each.
Fig. 4.Mean number ofwild recruits ofMya arenaria (SL≤15mm) and Carcinus maenas inexperimental units fromField Experiment I conducted atDBF (15 July to 29October 2011).(See Table 1 for a description of each level of the Exclusion treatment.) (n=5).
Table 7Analysis of variance on the arcsine-transformedmean percent survival of hatchery-rearedjuveniles of Mya arenaria from Field Experiment II at Duck Brook Flat, in Holmes Bay,Cutler, Maine (3 July to 28 October 2011). P-values in boldface represent statisticallysignificant hypothesis tests (α=0.05). (n=5).
69E.B.P. Tan, B.F. Beal / Journal of Experimental Marine Biology and Ecology 462 (2015) 62–73
units at the beginning of the study when ambient sediments were usedto fill units, and then became entrapped inside. With time, and success-ful molting or growth, these predators could have attained sizes largeenough to consume clams in an experimental unit; 3) Small predatorscould have crawled or recruited through the mesh aperture into theexperimental unit sometime during the field trial, then grown to asize where they became entrapped and, with continued growth, con-sumed the clams at some larger size; 4) The flexibility of some typesof plastic netting, combined with the mass of larger crabs, depressedthe netting to the sediment surface of the experimental unit permittingthe crustaceans an opportunity to insert the tips of their chelae throughthe net aperture and encounter the siphons or posterior ventral shellmargin of the clams, pulling them to the sediment surface then crushingthem through the netting, or using a mandibular chipping technique asdescribed by Morton and Harper (2008).
4.1. Green crab behavior in the field and laboratory
Although direct observation of mortality events did not take place inthe field, several lines of evidence suggest that individuals of Carcinusmaenas were primarily responsible for the loss of juveniles of Myaarenaria. For example, at the end of Field Experiment I & II, 10% and20% of the experimental units contained green crabs, respectively. Noother large decapods (e.g., Cancer irroratus – Beal, 2006a; Homarusamericanus) were found in or around experimental units at the end ofthe field trials in October 2011. In addition, green crab populationsalong the Maine coast have increased dramatically since 2009 (MEDMR, 2013; Whitlow and Grabowski, 2012; B. Beal, pers. obs.).
Table 6Analysis of variance on the untransformedmean number ofwild recruits (SL≤15 mm)ofMya aBrook Flat, in Holmes Bay, Cutler, Maine (15 July to 29 October 2011). P-values in boldface rep
Wild Mya arenaria recru
Source df SS MS
Exclusion 4 195.28 48.8200Fully Netted vs. Controls 1 144.21 144.21Unsupported vs. Supported Netting 1 40.02 40.02Unsupported (with vs. without metal rods) 1 1.25 1.25Collar vs. Roof Control 1 9.80 9.80
Crushed clams collected from experimental units in both the fieldand laboratory were typical of crustacean damage (Beal, 2006a) ratherthan from accidental damage. Valves of Mya that have been crushedby C. maenas, or other crustacean predators such as C. irroratus, have adistinctive chipping or crushing pattern (Boulding, 1984). Sometimesone valve is left nearly intact, while a fraction of the other remainsattached by the hinge ligament (this was the predominant damagetype observed in Laboratory Experiment II). In other instances, bothvalves are crushed nearly completely leaving only the umbo regionthat is held together by the hinge ligament (this damage type occurredmost often in the open controls [OEMC, OERC] in Field Experiment I andin Laboratory Experiment III). Previous field studies with juveniles ofM. arenaria have shown that live crabs can enter experimental units(either through the apertures as early post-larvae or from accidental in-clusion when ambient sediments are placed into units at the beginningof the trial) resulting inmass clammortality (Beal et al., 2001). This sce-nario occurred in a single experimental unit from Field Experiment IIwhen a live C. maenas (CW=30 mm) was discovered along with nolive clams, three disarticulated, crushed valves, and 22 of 24 individualsmissing.
Although few clams were found dead with crushed/chipped valvesor missing in the fully protected units from either field experiment(Tables 2 & 8), the three laboratory trials provide some resolutionabout how C. maenas may prey on juveniles of Mya in field enclosureswith predator deterrent netting present. In Laboratory Experiment I, asingle clam was discovered dead with chipped valves in the areawhere the siphons protrude near the posterior ventral margin. That ob-servation, and the fact that the flexible netting on that experimentalunit was closer to the sediment surface than it had been on previousdays, suggests that by sitting or resting on top of the netting, crabsmay depress the flexible netting enough to allow them to prey onclams underneath. Although the actual predation event was notobserved, we infer that crab behavior is similar to that outlined inhypothesis #4 (described above).
In Laboratory Experiment II, crabs ate voraciously, consuming allclams in two of the four LC units, leaving all valves chipped or crushed(DC); however, in another unit with a crab, all clams were found deadwith undamaged valves (DU; Table 11). The occurrence of dead clamswith undamaged valves in field settings has been associated with
renaria andmean number of green crabs, Carcinusmaenas, from Field Experiment I at Duckresent statistically significant hypothesis tests (α=0.05). (n=5).
Table 8Fate of juveniles of Mya arenaria in Field Experiment II at Duck Brook Flat, Cutler, Maine(3 July to 28 October 2011). Exclusion (“−” refers to open enclosures with a 1 cm stripofflexible netting [4.2 mmaperture] surrounding the periphery of each experimental unit[Area=0.0182 m2] to deter clammigration; “+” refers to experimental units fully coveredwith a piece of flexible netting [4.2 mm aperture] to deter predation). Means (±95% CI)are given for each fate category – %A, %DU, %DC, and %M, and these abbreviations aredescribed in Table 2. (n=5).
Table 10Analysis of variance on the untransformed mean number of microalgal cells × 104 mL−1
from Laboratory Experiment II (3–9 January 2012). All treatments included microalgaeinitially established at 5 × 104 cells mL−1 U−1: 1) control (C): no soft-shell clam juveniles,no crab; 2) crushed crab (CC): clams present, crab effluent; 3) live crab (LC): clams andcrab present; and, 4) clam control (CL): clams present, no crab. Cell counts were takenon days 3 (6 January) and 6 (9 January). Single degree-of-freedom orthogonal contrastsare presented for the treatment factor. P-values in boldface represent statistically signifi-cant hypothesis tests (α=0.05). (n=4).
LC & CC vs. C & CL 1 1313.2813 1313.2813 38.24 b0.0001LC vs. CC 1 30.2500 30.2500 0.88 0.3573C vs. CL 1 45.5625 45.5625 1.33 0.2607
Date X Treatment 3 124.0938 41.3646 1.20 0.3294Error 24 824.2500 34.3438Total 31 2469.4688
lls (x
104 ) m
L-1 +
95%
CI
15
20
25
30
Without CrabsWith Crabs
70 E.B.P. Tan, B.F. Beal / Journal of Experimental Marine Biology and Ecology 462 (2015) 62–73
non-predatory activity such as disease, mishandling, starvation (Beal,2006b; Micheli, 1997; Peterson, 1982), or from predators other thandecapods that typically leave no shell damage (Bourque et al., 2002;Cadée, 1994; Cha, 1994). If the probability of mishandling juveniles inthe experiment was 0.5 (clams either were mishandled or they werenot), then the probability that all six in one unit would have beenmishandled would be 0.0313 (2-tailed Binomial test). Similarly, diseaseseems unlikely, or more clams in the controls would havemet the samefate. For example, 3 of 48 clams in the control (CC, CL) units died duringthe six-day trial. If clams had died due to disease, then a reasonable es-timate of the proportion of diseased clams at the beginning of the trialwould be 3/48, or 0.0625. The probability that all six clams in the LCunit died due to disease would be b0.0001 (2-tailed Binomial test). Inaddition, clams likely did not starve because there was ample food forthem to eat (Fig. 5). Since the six DU clams were all within a unit con-taining a crab, all clamswere scattered on the surface and not discolored(as opposed to being burrowed in the sediment with both valvesdiscolored black, as was the fate of the three DU clams from the controlunits), and no tissue was present in the valves, it would appear that thecrab somehow consumed the clams without leaving any discernibleshell damage. Because green crabs are known to crush juvenile, infaunalbivalves (Ejdung et al., 2009; Ropes, 1968; this study), this result issurprising. Perhaps some individuals of C. maenas are able to consumejuveniles of Mya as individuals of the brachyuran crab, Hemigrapsussanguineus (25–40 mm CW), is known to consume larger M. arenaria(ca. 45 mm SL) (Brousseau et al., 2001). In that study, predation didnot involve shell crushing or chipping. Instead, crabs used their chelato pull tissue from the gaping siphonal (posterior) end of the clam,leaving the prey shells intact. C. maenas has been observed preying onblue mussels, Mytilus edulis, without inflicting shell damage, a methodtermed “boring” by Elner (1978). Other decapod crustaceans havebeen noted for their ability to prey on bivalves without inflicting shell
Table 9Percent frequency of the daily position of the green crab relative to experimental units(15 cm plastic pots) from 21 August to 11 September 2011 (22 days; Laboratory Experi-ment I). Each unit per aquarium was randomly assigned to one of three treatments:1) tight-fitting, flexible netting (i.e., without slack); 2) flexible netting with slack; and,3) rigid netting. Treatments 2 & 3 were similar to Exclusion treatment PENU in Field Ex-periment I (see Table 1). Percent of time crabs were burrowed also is presented.
damage (Lau, 1987). For most species of crabs, a critical bivalve preysize exists below which predation occurs (Seed and Hughes, 1995);however, in species with a mantle gape, as in Mya arenaria, crabs maybe able to cut the mantle tissue between the valves, sever at least oneadductormuscle and then begin to consume the tissueswithout damag-ing the shells. If C.maenas is able to prey on some juveniles ofMyawith-out leaving evidence of its attack in the shells, then predation rates bythis exotic predator on soft-shell clams may be higher than reportedelsewhere (e.g., Beal, 2006a,b; Hunt and Mullineaux, 2002; Whitlow,2010; Whitlow et al., 2003).
Results from Laboratory Experiment III provide the most conclu-sive evidence for the ability of green crabs to prey on juveniles ofM. arenaria that are purportedly protected with flexible, plasticnetting (aperture =6.4 mm) in our experimental units. Individualclams consumed by C. maenas occurred only in the treatment wherecrabs could access clams through the netting (High Sand; Table 12). Inthat laboratory trial, slack, flexible netting protected all clams, but be-cause sediment depth inside two of the three experimental units peraquarium was apparently lower than the depth to which the nettingcould stretch, clams in those units were safe compared to animalshoused in units filled with sediments. Beal and Kraus (2002) placedgroups of similar size experimental units containing cultured juvenilesof M. arenaria in ambient sediments at DBF, and surrounded themwith wooden boxes pushed into the sediments so that various types
Days0 1 2 3 4 5 6
Mea
n N
umbe
r of A
lgal
Ce
0
5
10
Fig. 5.Mean number of microalgal cells × 104mL−1 (Chaetoceros gracilis) from LaboratoryExperiment II (3–9 January 2012). Experimental unitswere stocked initiallywith 5.0 X 104
cells. Data from crushed crab (CC) and live crab (LC) treatments (“With Crabs”), and datafrom control (C) and clam control (CL) treatments (“Without Crabs”) were pooled due tothe lack of a significant difference between means (see Table 10) (n=4).
Table 11Mean number alive, dead uncrushed, dead crushed of juveniles of soft-shell clams and fi-nal salinity from Laboratory Experiment II (3–9 January 2012). C=control; CC=crushedcrab; LC=live crab; CL=clam control.
Carcinusmaenas
Mya arenaria
Block Treatment CW(mm)
Sex NumberAlive
NumberDU
NumberDC
Final Salinity(ppt)
1 C - - - - - 38CC 31.6 M 6 0 0 39LC 37.9 M 0 0 6 39CL - - 6 0 0 39
2 C - - - - - 36CC 38.6 F 6 0 0 38LC 32.9 M 6⁎ 0 0 36CL - - 6 0 0 36
3 C - - - - - 34CC 27.7 F 6 0 0 40LC 33.8 M 0 6⁎⁎ 0 40CL - - 5 1 0 35
4 C - - - - - 35CC 25.6 M 5 1 0 35LC 26.5 M 0 0 6⁎⁎⁎ 39CL - - 5 1 0 39
⁎ The crab was found burrowed, and was missing its right chelae.⁎⁎ These shells were articulated, had no signs of chipping, and had no tissue. They werefound on top of the sediment with their shells open less than 24 hours after the experi-ment was initiated.⁎⁎⁎ Of these six dead chipped clams, the valves of twowere articulated, and twoother clamswere missing half of one valve, but with the umbo intact.
71E.B.P. Tan, B.F. Beal / Journal of Experimental Marine Biology and Ecology 462 (2015) 62–73
of netting could be stretched tightly around the box to protect theclams. A gap of several cm existed between the netting and the units,and only a small percentage (b6 %) of clams were recovered deadwith crushed valves after 174 days in the field (April to October,1991). Similar results were observed by Spencer et al. (1992) who ex-amined the effectiveness of different types of protective netting on sur-vival of Tapes philippinarum in boxes on intertidal beaches near Conwyin Wales, UK.
4.2. Soft-shell clam responses to predators in the field and laboratory
Results from both field experiments underscore the relative impor-tance of predation, especially by decapods, in controlling populationsof juvenile soft-shell clams as others have shown (e.g., Beal, 2006b;
Table 12Number of live and dead crushed juveniles ofMya arenaria (xSL=14.8 ± 0.4 mm, n=30)in Laboratory Experiment III (27 June–17 July 2013). All individuals of C. maenaswere male. Each experimental unit in an aquarium was completely covered with a pieceof plastic, flexible netting (aperture=6.4 mm). ANOVA detected a significant treat-ment effect (P=0.0067) on the number of dead crushed clams per unit. Treatments:Low=experimental unit (15 cm diameter × 15 cm deep plastic horticultural pot)filled half-way with sand; Medium=unit filled with sand to within 1 cm of the rim;High=unit filled completely with sand to the rim.
Aquarium Crab CW (mm) Treatment Number of crushed clams
I 70.6 Low 0Medium 0High 0
II 70.3 Low 0Medium 0High 1
III 71.7 Low 0Medium 0High 2
IV 70.5 Low 0Medium 0High 1
V 69.9 Low 0Medium 0High 1
Beal et al., 2001; Floyd and Williams, 2004; Glude, 1955; Jensen andJensen, 1985; Smith et al., 1955; Welch, 1968; Whitlow et al., 2003).Combining field data from percent dead crushed/chipped and missingcategories (Tables 2 & 8) from open enclosures demonstrated thatlosses between 56.7–86.7% occurred in Field Experiment I, and up to87.5% in Field Experiment II. Missing clams are presumed dead (Beal,2006a), and together with animals identified as DC, these data demon-strate that the relative intensity of predation on small Mya at DBF ismuch higher than observed in the early 1990's over a similar time peri-od and in open enclosures (see Table 6 in Beal andKraus, 2002). Becausegreen crab population numbers respond directly to seawater tempera-tures (Welch, 1968), predation rates on soft-shell clams may haveincreased recently due to increases in crab numbers in response togradual warming of seawater in eastern Maine (Mills et al., 2013).
Number of 0-year class individuals ofMya (wild recruits) respondedsignificantly to predator deterrent netting in Field Experiment I(Table 6), but not in Field Experiment II. Examining all experimentalunits stocked initially with six clams in Experiment II (N=10), wild re-cruit densitywas approximately 4.5x greater in the protected (1.8± 1.8ind. unit−1, n=5) vs. control (0.4 ± 0.7 ind. unit−1, n=5) treatment(P=0.083). Enhancement of 0-year class individuals ofMya in predatorexclusion treatments in easternMaine has been shown to be variable inprevious studies. For example, Beal and Kraus (2002) observed signifi-cantly higher (3x) wild recruit densities in protected vs. control treat-ments at one intertidal site in eastern Maine but not at DBF. Beal et al.(2001) found no enhancement of wild recruits due to predator exclu-sion along a tidal gradient at an intertidal site approximately 20 kmwest of DBF; however, predator deterrence led to a 3-fold enhancementin numbers ofMya recruits at three of four intertidal sites in far easternMaine during 2003 (Beal, 2006b). In predator exclusion studiesconducted elsewhere, Mya responded with increased recruitment incages that protected small individuals from horseshoe crab attack inDelaware Bay, New Jersey (Bottom, 1984), from green crabs and fishin Barnstable Harbor, Massachusetts (Hunt and Mullineaux, 2002),and from C. maenas and Crangon crangon in the European WaddenSea, especially after mild winters (Strasser, 2002).
Laboratory Experiment II examined the feeding behavior of clamswith crabs present, with crab cues present, and in controls withoutcrabs or cues. Mean algal cell densities increased by 4x in treatmentswith crabs or cues compared to the controls; however, therewas no sig-nificant difference between controls with and without clams (P=0.26;Table 10). The power of this test was quite low (b0.50). Nonetheless, atthe end of the trial, mean algal cell density mL−1 in the control withoutclamswas 4.88± 3.07 × 104 (n=4), whichwas not significantly differ-ent than the initial stocking density of 5.0 × 104 cellsmL−1 (one-samplet-test [2-tailed], p=0.905). Conversely,finalmeandensity in the controlwith clams was 3.25 ± 0.79 × 104 mL−1 (n=4), and this was signifi-cantly different from the initial stocking density (P=0.006). This sug-gests that clams in the units with only the microalgae consumed thecultured phytoplankton over the 6-day trial, but clams in the live crabtreatment (LC) that survived (one of four units; Table 11) and clamsin the units with the crab cues (CC) did not feed at all, or consumedless phytoplankton than conspecifics in the CL treatment. Without thebenefit of a crushed crab control (no clams plus crab due), anotherexplanation is that excess nutrients (e.g., nitrogen and its metabolites)associated with the LC and CC treatments enhanced algal growth, andlive clams in those treatments actually fed on algae at rates similar toclams in the CL treatment. In other lab trials, soft-shell clams increasedtheir burial depth when activity simulating the red rock crab, Cancerproductus, occurred (Zaklan and Ydenberg, 1997). In the field,Whitlow et al. (2003) were able to induce a similar behavior whenadults of Mya burrowed ca. 12% deeper in plots that included greencrabs vs. plots where crabs were excluded. Similar results were ob-served by Whitlow (2010) in the laboratory; however, no differencesin shell growth were detected between treatments with and withoutcrabs or their cues. Beal et al. (2001) found that shell growth of juveniles
72 E.B.P. Tan, B.F. Beal / Journal of Experimental Marine Biology and Ecology 462 (2015) 62–73
of Mya was 6.6% lower during April to December in experimental fieldunits in eastern Maine intertidal flat when predators (moon snails,Lunatia heros, or green crabs) were unintentionally included inexclosure units than when they were not. The data presented here, incombination with other independent trials and observations, suggestthat green crabs can induce both direct and indirect responses in soft-shell clam behavior that may be size- or age-dependent. Hard clams,Mercenaria mercenaria, behaved similarly in the lab (i.e., reduced feed-ing in the presence of predators – knobbed whelks, Busycon carica;blue crabs, Callinectes sapidus) (Smee and Weissburg, 2006). In fieldenclosures, long-term exposure to predators (whelks) resulted indepressed hard clam growth rates compared to controls (Nakaoka,2000). Together, these studies and the laboratory trial presented here,suggest that these bivalves can alter their behavior by reducing theirchemical presence in an attempt to decrease their susceptibility tomobile predators that use sensory cues from prey while foraging.
5. Conclusion
The invasive green crab, Carcinus maenas, is an ecosystem engineerthat alters the function and organization of marine habitats (intertidalrocky shores – see Trussell et al., 2003; Spartina marshes – seeMcDonald et al., 2006; Ropes, 1968; eelgrass beds – see Davis et al.,1998; Garbary and Miller, 2006; intertidal mudflats – see Grosholzet al., 2000; and subtidal locations – see Elner, 1981; Donahue et al.,2009), reduces biodiversity and alters food webs (Garbary et al., 2014;Pejchar and Mooney, 2009). Green crabs are omnivores (Baeta et al.,2006; Griffen, 2014), and their diet reflects their size and habitat(Elner and Hughes, 1978; Grosholz and Ruiz, 1996; Mascaro and Seed,2001; Rangeley and Thomas, 1987; Sungail et al., 2013), but they preferbivalvemolluscs (Elner, 1981; Grosholz andRuiz, 1995; Ropes, 1968). IneasternMaine, USA, green crabs are amajor threat to bothwild and cul-tured populations of Mya arenaria (Beal and Kraus, 2002; Beal et al.,2001). Green crabs attack juveniles ofM. arenaria in the field and labo-ratory using a variety of tactics that can result in typical shell damageranging from ventral margin chipping to complete breakage of bothvalves leaving only the umbos intact. In addition, some crabs apparentlycan consume soft-shell clamswithout inflicting any damage to the shell.Whether this occurs by severing the mantle tissue that gapes betweenvalves of this infaunal bivalve or by othermeans is unclear, but suggeststhat previous investigations assessing the relationship between greencrabs and their soft-shell clam prey (Beal, 2006a; Floyd and Williams,2004; Flynn and Smee, 2010) may have underestimated the effect thispredator plays in the early life-history of this bivalve.
Funding
This work was supported by the University of Maine at Machias andthe Downeast Institute for Applied Marine Research and Education.
Acknowledgments
We thank Downeast Institute staff C. Jourdet, K. Pepperman, and G.Protopopescu for providing clams and logistical support for LaboratoryExperiments I & II. In addition, the following University of Maine atMachias (UMM) students assisted us in the field with various aspectsof Experiments I & II, which we are grateful: J. Berninger, L. Frauen,S. Lesnock, J. Parlin, and S. Stoddard. This effort was part of a seniorthesis for E. Tan through UMM. [SS].
References
Babcock, R.C., Kelly, S., Shears, N.T., Walker, J.W., Willis, T.J., 1999. Changes in communitystructure in temperate marine reserves. Mar. Ecol. Prog. Ser. 189, 125–134.
Baeta, A., Cabral, H.N., Marques, J.O., Pardal, M.A., 2006. Feeding ecology of the greencrab, Carcinus maenas (L., 1758) in a temperate estuary, Portugal. Crustaceana 79,1181–1193.
Beal, B.F., 1983. Predation of juveniles of the hard clam Mercenaria mercenaria (Linne)by the snapping shrimp Alpheus heterochaelis Say and Alpheus normanni Kingsley.J. Shellfish Res. 3, 1–9.
Beal, B.F., 2002. Adding value to live, commercial size soft-shell clams (Mya arenaria L.) inMaine, USA: results from repeated, small-scale, field impoundment trials. Aquacul-ture 210, 119–135.
Beal, B.F., 2006a. Biotic and abiotic factors influencing growth and survival of wild andcultured individuals of the softshell clam (Mya arenaria L.) in Eastern Maine. J. Shell-fish Res. 25, 461–474.
Beal, B.F., 2006b. Relative importance of predation and intraspecific competition inregulating growth and survival of juveniles of the soft-shell clam, Mya arenaria L.,at several spatial scales. J. Exp. Mar. Biol. Ecol. 336, 1–17.
Beal, B.F., Kraus, M.G., 2002. Interactive effects of initial size, stocking density, and typeof predator deterrent netting on survival and growth of cultured juveniles of thesoft-shell clam, Mya arenaria L., in eastern Maine. Aquaculture 208, 81–111.
Beal, B.F., Lithgow, C.D., Shaw, D.P., Renshaw, S., Ouellette, D., 1995. Overwinteringhatchery-reared individuals of the soft-shell clam, Mya arenaria L.: a field test ofsite, clam size, and intraspecific density. Aquaculture 130, 145–158.
Beal, B.F., Bayer, R., Kraus, M.G., Chapman, S.R., 1999. A unique shell marker in juvenile,hatchery–reared individuals of the softshell clam, Mya arenaria L. Fish. Bull. 97,380–386.
Beal, B.F., Parker, M.R., Vencile, K.W., 2001. Seasonal effects of intraspecific density andpredator exclusion along a shore-level gradient on survival and growth of juvenilesof the soft-shell clam, Mya arenaria L., in Maine, USA. J. Exp. Mar. Biol. Ecol. 264,133–169.
Beukema, J.J., 1976. Biomass and species richness of the macro-benthic animals living onthe tidal flats of the Dutch Wadden Sea. Neth. J. Sea Res. 10, 236–261.
Blundon, J.A., Kennedy, V.S., 1982. Mechanical and behavioral aspects of blue crab,Callinectes sapidus (Rathbun), predation on Chesapeake Bay bivalves. J. Exp. Mar.Biol. Ecol. 65, 47–65.
Bottom, M.L., 1984. The importance of predation by horseshoe crabs, Limulus polyphemus,to an intertidal sand flat community. J. Mar. Res. 42, 139–161.
Boulding, E.G., 1984. Crab-resistant features of shells of burrowing bivalves: decreasedvulnerability by increasing handling time. J. Exp. Mar. Biol. Ecol. 76, 201–223.
Bourque, D., Miron, G., Landry, T., 2001. Predation on soft-shell clams (Mya arenaria) bythe nemertean Cerebratulus lacteus in Atlantic Canada: implications for controlmeasures. Hydrobiologia 456, 33–44.
Bourque, D., Miron, G., Landry, T., 2002. Predatory relationship between the nemerteanCerebratulus lacteus and the soft-shell clam, Mya arenaria: surface-exploration activ-ity and qualitative observations on feeding behaviour. Can. J. Zool. 80, 1204–1211.
Bowen, J.E., Hunt, H.L., 2009. Settlement and recruitment patterns of the soft-shell clam,Mya arenaria, on the northern shore of the Bay of Fundy, Canada. Estuar. Coast. 32,758–772.
Brousseau, D.J., Filipowicz, A., Baglivo, J.A., 2001. Laboratory investigations of the effectsof predator sex and size on prey selection by the Asian crab, Hemigrapsus sanguineus.J. Exp. Mar. Biol. Ecol. 262, 199–210.
Cadée, G.C., 1994. Eider, shelduck, and other predators, the main producers of shellfragments in the Wadden Sea: palaeoecological implications. Palaeontology 37,181–202.
Cha, M.W., 1994. Predation impacts on the population structure of infaunal bivalves fromLobster Bay, Cape D'Aguilar, Hong Kong. In: Morton, B. (Ed.), The Malacofauna ofHong Kong and Southern China IIIProceedings of the Third International Workshopon the Malacofauna of Hong Kong and Southern China, Hong Kong 13 April - 1May1992. Hong Kong Univ. Press, Hong Kong, pp. 391–404.
Cigarría, J., Fernández, J.M., 2000. Management of Manila clam beds I. Influence ofseed size, type of substratum and protection on initial mortality. Aquaculture 182,173–182.
Cohen, A.N., Carlton, J.T., Fountain, M.C., 1995. Introduction, dispersal and potentialimpacts of the green crab Carcinus maenas in San Francisco Bay, California. Mar.Biol. 122, 225–237.
Commito, J.A., 1982. Effects of Lunatia heros predation on the population dynamics ofMyaarenaria and Macoma balthica in Maine, USA. Mar. Biol. 69, 187–193.
Conde, A., Novais, J., Dominguez, J., 2010. Southern limit of distribution of the soft-shellclam Mya arenaria on the Atlantic east coast. Biol. Invasions 12, 429–432.
Davis, R.C., Short, F.T., Burdick, D.M., 1998. Quantifying the effects of green crab damage toeelgrass transplants. Restor. Ecol. 6, 297–302.
Donahue, M.J., Nichols, A., Santamaria, C.A., League-Pike, P.E., Krediet, C.J., Perez, K.O.,Shulman, M.J., 2009. Predation risk, prey abundance, and the vertical distribution ofthree brachyuran crabs on Gulf of Maine shores. J. Crustac. Biol. 29, 523–531.
Dow, R.L., 1977. Effects of climatic cycles on the relative abundance and availability ofcommercial marine and estuarine species. J. Cons. Int. Explor. Mer 37, 274–280.
Edwards, D.C., Hubner, J.D., 1977. Feeding and growth rates of Polinices duplicatus preyingon Mya arenaria at Barnstable Harbor, Massachusetts. Ecology 58, 1218–1236.
Ejdung, G., Flach, E., Byrén, L., Hummel, H., 2009. Predation by crustaceans on native andnon-native Baltic clams. Aquat. Biol. 6, 15–24.
Elner, R.W., 1978. The mechanics of predation by the shore crab, Carcinus maenas (L.), onthe edible mussel, Mytilus edulis L. Oecologia 36, 333–344.
Elner, R.W., 1981. Diet of green crab Carcinus maenas (L.) from Port Herbert, southwesternNova Scotia. J. Shellfish Res. 1, 89–94.
Elner, R.W., Hughes, R.N., 1978. Energy maximization in the diet of the shore crab,Carcinus maenas. J. Anim. Ecol. 47, 103–116.
Floyd, T., Williams, J., 2004. Impact of green crab (Carcinus maenas L.) predation on a pop-ulation of softshell clams (Mya arenaria L.) in the Southern Gulf of St. Lawrence. J.Shellfish Res. 23, 457–462.
73E.B.P. Tan, B.F. Beal / Journal of Experimental Marine Biology and Ecology 462 (2015) 62–73
Flynn, A.M., Smee, D.L., 2010. Behavioral plasticity of the soft-shell clam,Mya arenaria (L.), inthe presence of predators increases survival in the field. J. Exp. Mar. Biol. Ecol. 383,32–38.
Freestone, A.L., Osman, R.W., Ruiz, G.M., Torchin, M.E., 2011. Stronger predation in the tro-pics shapes species richness patterns in marine communities. Ecology 92, 983–993.
Garbary, D.J., Miller, A.G., 2006. Green crabs (Carcinus maenas) as the grim reaper:Destruction of eelgrass beds in Nova Scotia. J. Shellfish Res. 25, 728.
Garbary, D.J., Miller, A.G., Williams, J., Seymour, N.R., 2014. Drastic decline of an extensiveeelgrass bed in Nova Scotia due to the activity of the invasive green crab (Carcinusmaenas). Mar. Biol. 161, 3–15.
Glude, J.B., 1955. The effects of temperature and predators on the abundance of the soft-shell clam, Mya arenaria, in New England. Trans. Am. Fish. Soc. 84, 13–26.
Griffen, B.D., 2014. Linking individual diet variation and fecundity in an omnivorousmarine consumer. Oecologia 174, 121–130.
Grosholz, E.D., Ruiz, G.M., 1995. Spread and potential impact of the recently introducedEuropean green crab, Carcinus maenas, in central California. Mar. Biol. 122, 239–247.
Grosholz, E.D., Ruiz, G.M., 1996. Predicting the impact of introduced marine species:lessons from the multiple invasions of the European green crab, Carcinus maenas.Biol. Conserv. 78, 59–66.
Grosholz, E.D., Ruiz, G.M., Dean, C.A., Shirley, K.A., Maron, J.L., Connors, P.G., 2000. The im-pacts of a nonindigenous marine predator in a California bay. Ecology 81, 1206–1224.
Hunt, H., 2004. Effects of epibenthic predators in flow: transport and mortality ofjuveniles of the soft shell clam Mya arenaria. Mar. Ecol. Prog. Ser. 279, 151–160.
Hunt, H., Mullineaux, L.S., 2002. The roles of predation and postlarval transport in recruit-ment of the soft shell clam (Mya arenaria). Limnol. Oceanogr. 47, 151–164.
Irlandi, E.A., 1997. Seagrass patch size and survivorship of an infaunal bivalve. Oikos 78,511–518.
Jensen, K.T., Jensen, J.N., 1985. The importance of some epibenthic predators on thedensity of juvenile benthic macrofauna in the Danish Wadden Sea. J. Exp. Mar. Biol.Ecol. 89, 157–174.
Kelso, W.E., 1979. Predation on soft-shell clams, Mya arenaria, by the common mummi-chog, Fundulus heteroclitus. Estuaries 2, 249–254.
Kuhlmann, M.L., Hines, A.H., 2005. Density-dependent predation by blue crabs Callinectessapidus on natural prey populations of infaunal bivalves. Mar. Ecol. Prog. Ser. 295,215–228.
Lau, C.J., 1987. Feeding behaviour of the Hawaiian slipper lobster, Scyllaridessquammousus, with a review of decapods crustacean feeding tactics on molluscanprey. Bull. Mar. Sci. 41, 378–391.
LeBlanc, S., Miron, G., 2006. Bentho-pelagic distribution of early stages of softshell clams(Mya arenaria) in tidally contrasted regimes. Can. J. Zool. 84, 459–472.
Maine Department of Marine Resources, 2013. Green crabs in Maine: An overview – July2013. http://www.maine.gov/dmr/rm/invasives/GreenCrabs.htm.
Mascaro, M., Seed, R., 2001. Foraging behavior of juvenile Carcinus maenas (L.) and Cancerpagurus L. Mar. Biol. 139, 1135–1145.
McDonald, P.S., Holsman, K.K., Beauchamp, D.A., Dumbauld, B.R., Armstrong, D.A., 2006.Bioenergetics modeling to investigate habitat use by the nonindigenous crab,Carcinus maenas, in Willipa Bay, Washington. Estuar. Coast. 29, 1132–1149.
Micheli, F., 1997. Effects of predator foraging behavior on patterns of prey mortality inmarine soft bottoms. Ecol. Monogr. 67, 203–224.
Mills, K.E., Pershing, A.J., Brown, C.J., Chen, Y., Chiang, F.-S., Holland, D.S., Lehuta, S., Nye, J.A.,Sun, J.C., Thomas, A.C., Wahle, R.A., 2013. Fisheries management in a changing climate:Lessons from the 2012 heatwave in the Northwest Atlantic. Oceanography 26, 191–195.
Morton, B., Harper, E.M., 2008. Predation upon Mytilus galloprovincialis (Mollusca:Bivalvia: Mytilidae) by juvenile Carcinus maenas (Crustacea: Decapoda) usingmandibular chipping. J. Mar. Biol. Assoc. U. K. 88, 563–568.
Nakaoka, M., 2000. Nonlethal effects of predators on prey populations: predator-mediated change in bivalve growth. Ecology 81, 1031–1045.
Pejchar, L., Mooney, H.A., 2009. Invasive species, ecosystem services and human well-being. Trends Ecol. Evol. 24, 497–504.
Peterson, C.H., 1982. Clam predation by whelks (Busycon spp.): experimental tests of theimportance of prey size, prey density, and seagrass cover. Mar. Biol. 66, 159–170.
Peterson, C.H., Black, R., 1993. Experimental tests of the advantages and disadvantages ofhigh density for two coexisting cockles in a Southern Ocean lagoon. J. Anim. Ecol. 62,614–633.
Pickering, T., Quijón, P.A., 2011. Potential effects of a non-indigenous predator in itsexpanded range: assessing green crab, Carcinus maenas, prey preference in a produc-tive coastal area of Atlantic Canada. Mar. Biol. 158, 2065–2078.
Rangeley, R.W., Thomas, M.L.H., 1987. Predatory behavior of juvenile shore crab Carcinusmaenas (L.). J. Exp. Mar. Biol. Ecol. 108, 191–197.
Ropes, J.W., 1968. The feeding habits of the green crab, Carcinus maenas (L.). Fish. Bull. 67,183–203.
Sallan, L.C., Kammer, T.W., Ausich, W.I., Cook, L.A., 2011. Persistent predator–preydynamics revealed by mass extinction. Proc. Natl. Acad. Sci. 108, 8335–8338.
Seed, R., Hughes, R.N., 1995. Criteria for prey size-selection in molluscivorous crabs withcontrasting claw morphologies. J. Exp. Mar. Biol. Ecol. 193, 177–195.
Seitz, R.D., Lipcius, R.M., Hines, A.H., Eggleston, D.B., 2001. Density-dependent predation,habitat variation, and the persistence of marine bivalve prey. Ecology 82, 2435–2451.
Seitz, R.D., Marshall Jr., L.S., Hines, A.H., Clark, K.L., 2003. Effects of hypoxia on predator–prey dynamics of the blue crab Callinectes sapidus and the Baltic clam Macomabalthica in Chesapeake Bay. Mar. Ecol. Prog. Ser. 257, 179–188.
Serdar, S., Lök, A., Köse, A., Yildiz, H., Acarli, S., Goulletquer, P., 2007. Growth and survivalrates of carpet shell clam (Tapes decussatus Linnaeus, 1758) using various culturemethods in Sufa (Homa) Lagoon, Izmir, Turkey. Aquacult. Eng. 37, 89–99.
Shaw, R.G., Mitchell-Olds, T., 1993. ANOVA for unbalanced data: an overview. Ecology 74,1638–1645.
Skilleter, G.A., 1994. Refuges from predation and the persistence of estuarine clampopulations. Mar. Ecol. Prog. Ser. 109, 29–42.
Smith, L.D., Hines, A.H., 1991. The effect of cheliped loss on blue crab Callinectes sapidusRathbun foraging rate on soft-shell clams Mya arenaria L. J. Exp. Mar. Biol. Ecol.151, 245–256.
Smith, O.R., Baptist, J.P., Chin, E., 1955. Experimental farming of the soft-shell clam, Myaarenaria, in Massachusetts, 1949–1953. U.S. Fish Wildl. Serv. Commer. Fish. Rev. 17,1–16.
Smith, T.E., Ydenberg, R.C., Elner, R.W., 1999. Foraging behavior of an excavating predator,the red rock crab (Cancer productus Randall) on soft-shell clam (Mya arenaria L.). J.Exp. Mar. Biol. Ecol. 238, 185–197.
Stanley, S.M., 2008. Predation defeats competition on the seafloor. Paleobiology 34, 1–21.Steimle, F.W., Pikanowski, R.A., McMillan, D.G., Zetlin, C.A., Wilk, S.J., 2000. Demersal fish
and American lobster diets in the Lower Hudson – Raritan Estuary. NOAA Tech. Mem.NMFS-NE-161, (118 pp., http://nefsc.noaa.gov/publications/tm/tm161/tm161.pdf).
Strasser, M., 2002. Reduced epibenthic predation on intertidal bivalves after a severewinter in the European Wadden Sea. Mar. Ecol. Prog. Ser. 241, 113–123.
Sungail, J., Brown, A.C., Alpert, K., Maurukus, J., 2013. Prey selection by Gulf of Maine greencrabs (Carinus maenas), rock crabs (Caner irroratus), and American lobsters (Homarusamericanus): A laboratory study. J. Exp. Mar. Biol. Ecol. 449, 294–303.
Taylor, D.L., Eggleston, D.B., 2000. Effects of hypoxia on an estuarine predator–preyinteraction: foraging behavior and mutual interference in the blue crab Callinectessapidus and the infaunal clam prey Mya arenaria. Mar. Ecol. Prog. Ser. 196, 221–237.
Underwood, A.J., 1997. Experiments in Ecology: Their Logical Design and InterpretationUsing Analysis of Variance. Cambridge Univ. Press, Cambridge, UK, pp. 364–365.
Weinberg, J.R., Leavitt, D.F., Lancaster, B.A., McDowell Capuzzo, J., 1997. Experimental fieldstudies with Mya arenaria (Bivalvia) on the induction and effect of hematopoieticneoplasia. J. Invertebr. Pathol. 69, 183–194.
Welch, W.R., 1968. Changes in abundance of the green crab, Carcinus maenas (L.), inrelation to recent temperature changes. Fish. Bull. 67, 337–345.
Whitlow, W.L., 2010. Changes in survivorship, behavior, and morphology in native soft-shell clams induced by invasive green crab predators. Mar. Ecol. 31, 418–430.
Whitlow, W.L., Grabowski, J.H., 2012. Examining how landscapes influenced benthiccommunity assemblages in seagrass and mudflat habitats in southern Maine. J. Exp.Mar. Biol. Ecol. 411, 1–6.
Whitlow, W.L., Rice, N.A., Sweeney, C., 2003. Native species vulnerability to introducedpredators: testing an inducible defense and a refuge from predation. Biol. Invasions5, 23–31.
Winer, B.J., Brown, D.R., Michels, K.M., 1991. Statistical principles in experimental design,3rd ed. McGraw-Hill, New York (1057 pp.).
Wong, M.C., Peterson, C.H., Kay, J., 2010. Prey size selection and bottom type influencemultiple predator effects in a crab-bivalve system. Mar. Ecol. Prog. Ser. 409, 143–156.
Zaklan, S.D., Ydenberg, R., 1997. The body size-burial depth relationship in the infaunalclam Mya arenaria. J. Exp. Mar. Biol. Ecol. 215, 1–17.
![AP* Biology: Ecology Practice MC [Version Map] Biology Ecology... · AP* Biology: Ecology Practice MC [Version Map] 1 ABCD MC 1 8 9 7 ... AP* Biology: Ecology Practice MC ... which](https://static.documents.pub/doc/80x56/5b449d207f8b9ae0668bd35b/ap-biology-ecology-practice-mc-version-map-biology-ecology-ap-biology.jpg)
